Are hadrons simpler than they seem?

I briefly review a systematic approximation scheme of QCD in which the quark model picture of hadrons emerges at lowest order. A linear A^0 potential arises if Gauss' law is solved with a non-vanishing boundary condition at spatial infinity. Similarly to the Dirac case one can describe relativistic states including any number of particle pairs (sea quarks) using valence wave functions, whose norms give {\em inclusive} probability densities. Provided \alpha_s(Q^2) freezes in the infrared, perturbative corrections to the S-matrix can be calculated in the usual way, but with states bound by the linear \order{\alpha_s^0} potential instead of plane waves in the in and out states.